Training in progress, step 282, checkpoint

checkpoint-282/added_tokens.json

@@ -0,0 +1,5 @@
+{
+  "<|endofprompt|>": 100276,
+  "<|im_end|>": 100265,
+  "<|im_start|>": 100264
+}

checkpoint-282/chat_template.jinja

@@ -0,0 +1,4 @@
+{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% for message in messages %}{{'<|im_start|>' + message['role'] + '
+' + message['content'] + '<|im_end|>' + '
+'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant
+' }}{% endif %}

checkpoint-282/config.json

@@ -0,0 +1,32 @@
+{
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 100257,
+  "embd_pdrop": 0.0,
+  "eos_token_id": 100265,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 8192,
+  "max_position_embeddings": 262144,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 8,
+  "pad_token_id": 100257,
+  "pretraining_tp": 1,
+  "resid_pdrop": 0.0,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 128000000,
+  "tie_word_embeddings": true,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.52.4",
+  "use_cache": false,
+  "vocab_size": 102400
+}

checkpoint-282/generation_config.json

@@ -0,0 +1,9 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 100257,
+  "do_sample": true,
+  "eos_token_id": 100265,
+  "pad_token_id": 100257,
+  "transformers_version": "4.52.4",
+  "use_cache": false
+}

checkpoint-282/model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:99044790e2d064bf8559a16d99a20473efab74cb5a047fc445e142b91f1d4d24
+size 4983077832

checkpoint-282/model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:18db60ae16f6f83b77c2bb84a7550f81fe0607f1fbb3761f2940c96b2084298c
+size 3842234168

checkpoint-282/model.safetensors.index.json

@@ -0,0 +1,261 @@
+{
+  "metadata": {
+    "total_size": 8825282560
+  },
+  "weight_map": {
+    "model.embed_tokens.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.norm.weight": "model-00002-of-00002.safetensors"
+  }
+}

checkpoint-282/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d0985d2fee9eaf37acf0486d816ea2576408d6e44e654694fecb6e5603a8b736
+size 11466963134

checkpoint-282/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b1a97db8e41139aa1239ba7fb79ddeb0af5998c6305a440c1fe182e6ad02f2f5
+size 14244

checkpoint-282/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e6574c450b3dc19a94f903b026d84caab58438abeecc47a4b4bbb8468d394fc
+size 1064

checkpoint-282/special_tokens_map.json

@@ -0,0 +1,30 @@
+{
+  "bos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-282/tokenizer_config.json

@@ -0,0 +1,69 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "100257": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100258": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100259": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100260": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100264": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "100265": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100276": {
+      "content": "<|endofprompt|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|endoftext|>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "extra_special_tokens": {},
+  "model_max_length": 262144,
+  "pad_token": "<|endoftext|>",
+  "tokenizer_class": "GPT2Tokenizer",
+  "unk_token": "<|endoftext|>"
+}

checkpoint-282/trainer_state.json

@@ -0,0 +1,2008 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.989399293286219,
+  "eval_steps": 500,
+  "global_step": 282,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.007067137809187279,
+      "grad_norm": 1.2265625,
+      "learning_rate": 0.0,
+      "loss": 1.9174,
+      "step": 1
+    },
+    {
+      "epoch": 0.014134275618374558,
+      "grad_norm": 1.0234375,
+      "learning_rate": 3.5714285714285716e-07,
+      "loss": 2.2074,
+      "step": 2
+    },
+    {
+      "epoch": 0.02120141342756184,
+      "grad_norm": 1.2421875,
+      "learning_rate": 7.142857142857143e-07,
+      "loss": 1.7854,
+      "step": 3
+    },
+    {
+      "epoch": 0.028268551236749116,
+      "grad_norm": 1.1484375,
+      "learning_rate": 1.0714285714285714e-06,
+      "loss": 2.1768,
+      "step": 4
+    },
+    {
+      "epoch": 0.0353356890459364,
+      "grad_norm": 0.9375,
+      "learning_rate": 1.4285714285714286e-06,
+      "loss": 2.1438,
+      "step": 5
+    },
+    {
+      "epoch": 0.04240282685512368,
+      "grad_norm": 3.484375,
+      "learning_rate": 1.7857142857142859e-06,
+      "loss": 2.5976,
+      "step": 6
+    },
+    {
+      "epoch": 0.04946996466431095,
+      "grad_norm": 1.2265625,
+      "learning_rate": 2.1428571428571427e-06,
+      "loss": 2.2351,
+      "step": 7
+    },
+    {
+      "epoch": 0.05653710247349823,
+      "grad_norm": 0.984375,
+      "learning_rate": 2.5e-06,
+      "loss": 2.0092,
+      "step": 8
+    },
+    {
+      "epoch": 0.0636042402826855,
+      "grad_norm": 1.1015625,
+      "learning_rate": 2.8571428571428573e-06,
+      "loss": 2.1347,
+      "step": 9
+    },
+    {
+      "epoch": 0.0706713780918728,
+      "grad_norm": 1.03125,
+      "learning_rate": 3.2142857142857147e-06,
+      "loss": 2.0748,
+      "step": 10
+    },
+    {
+      "epoch": 0.07773851590106007,
+      "grad_norm": 0.9140625,
+      "learning_rate": 3.5714285714285718e-06,
+      "loss": 1.7444,
+      "step": 11
+    },
+    {
+      "epoch": 0.08480565371024736,
+      "grad_norm": 0.83984375,
+      "learning_rate": 3.928571428571429e-06,
+      "loss": 1.5372,
+      "step": 12
+    },
+    {
+      "epoch": 0.09187279151943463,
+      "grad_norm": 0.71484375,
+      "learning_rate": 4.2857142857142855e-06,
+      "loss": 1.4944,
+      "step": 13
+    },
+    {
+      "epoch": 0.0989399293286219,
+      "grad_norm": 0.8046875,
+      "learning_rate": 4.642857142857144e-06,
+      "loss": 1.8555,
+      "step": 14
+    },
+    {
+      "epoch": 0.10600706713780919,
+      "grad_norm": 0.92578125,
+      "learning_rate": 5e-06,
+      "loss": 1.5569,
+      "step": 15
+    },
+    {
+      "epoch": 0.11307420494699646,
+      "grad_norm": 1.34375,
+      "learning_rate": 5.357142857142857e-06,
+      "loss": 2.1602,
+      "step": 16
+    },
+    {
+      "epoch": 0.12014134275618374,
+      "grad_norm": 0.84375,
+      "learning_rate": 5.7142857142857145e-06,
+      "loss": 2.0495,
+      "step": 17
+    },
+    {
+      "epoch": 0.127208480565371,
+      "grad_norm": 2.9375,
+      "learning_rate": 6.071428571428571e-06,
+      "loss": 2.2094,
+      "step": 18
+    },
+    {
+      "epoch": 0.13427561837455831,
+      "grad_norm": 0.8046875,
+      "learning_rate": 6.4285714285714295e-06,
+      "loss": 1.3931,
+      "step": 19
+    },
+    {
+      "epoch": 0.1413427561837456,
+      "grad_norm": 0.71875,
+      "learning_rate": 6.785714285714287e-06,
+      "loss": 1.7009,
+      "step": 20
+    },
+    {
+      "epoch": 0.14840989399293286,
+      "grad_norm": 0.859375,
+      "learning_rate": 7.1428571428571436e-06,
+      "loss": 1.8362,
+      "step": 21
+    },
+    {
+      "epoch": 0.15547703180212014,
+      "grad_norm": 0.67578125,
+      "learning_rate": 7.500000000000001e-06,
+      "loss": 1.9375,
+      "step": 22
+    },
+    {
+      "epoch": 0.1625441696113074,
+      "grad_norm": 0.9765625,
+      "learning_rate": 7.857142857142858e-06,
+      "loss": 2.1988,
+      "step": 23
+    },
+    {
+      "epoch": 0.1696113074204947,
+      "grad_norm": 0.71484375,
+      "learning_rate": 8.214285714285714e-06,
+      "loss": 1.7918,
+      "step": 24
+    },
+    {
+      "epoch": 0.17667844522968199,
+      "grad_norm": 1.0234375,
+      "learning_rate": 8.571428571428571e-06,
+      "loss": 2.1915,
+      "step": 25
+    },
+    {
+      "epoch": 0.18374558303886926,
+      "grad_norm": 0.796875,
+      "learning_rate": 8.92857142857143e-06,
+      "loss": 2.3101,
+      "step": 26
+    },
+    {
+      "epoch": 0.19081272084805653,
+      "grad_norm": 0.71875,
+      "learning_rate": 9.285714285714288e-06,
+      "loss": 1.9351,
+      "step": 27
+    },
+    {
+      "epoch": 0.1978798586572438,
+      "grad_norm": 1.4296875,
+      "learning_rate": 9.642857142857144e-06,
+      "loss": 1.7059,
+      "step": 28
+    },
+    {
+      "epoch": 0.2049469964664311,
+      "grad_norm": 0.7734375,
+      "learning_rate": 1e-05,
+      "loss": 2.1424,
+      "step": 29
+    },
+    {
+      "epoch": 0.21201413427561838,
+      "grad_norm": 0.6328125,
+      "learning_rate": 9.999914116658896e-06,
+      "loss": 1.7643,
+      "step": 30
+    },
+    {
+      "epoch": 0.21908127208480566,
+      "grad_norm": 0.79296875,
+      "learning_rate": 9.999656469585957e-06,
+      "loss": 1.8743,
+      "step": 31
+    },
+    {
+      "epoch": 0.22614840989399293,
+      "grad_norm": 0.74609375,
+      "learning_rate": 9.999227067632223e-06,
+      "loss": 2.0273,
+      "step": 32
+    },
+    {
+      "epoch": 0.2332155477031802,
+      "grad_norm": 0.8359375,
+      "learning_rate": 9.99862592554908e-06,
+      "loss": 1.8072,
+      "step": 33
+    },
+    {
+      "epoch": 0.24028268551236748,
+      "grad_norm": 0.82421875,
+      "learning_rate": 9.997853063987768e-06,
+      "loss": 2.0589,
+      "step": 34
+    },
+    {
+      "epoch": 0.24734982332155478,
+      "grad_norm": 0.66796875,
+      "learning_rate": 9.996908509498662e-06,
+      "loss": 1.9282,
+      "step": 35
+    },
+    {
+      "epoch": 0.254416961130742,
+      "grad_norm": 1.046875,
+      "learning_rate": 9.995792294530356e-06,
+      "loss": 1.538,
+      "step": 36
+    },
+    {
+      "epoch": 0.26148409893992935,
+      "grad_norm": 0.81640625,
+      "learning_rate": 9.994504457428557e-06,
+      "loss": 1.9843,
+      "step": 37
+    },
+    {
+      "epoch": 0.26855123674911663,
+      "grad_norm": 1.84375,
+      "learning_rate": 9.993045042434772e-06,
+      "loss": 2.7539,
+      "step": 38
+    },
+    {
+      "epoch": 0.2756183745583039,
+      "grad_norm": 0.66796875,
+      "learning_rate": 9.99141409968477e-06,
+      "loss": 1.7211,
+      "step": 39
+    },
+    {
+      "epoch": 0.2826855123674912,
+      "grad_norm": 0.71875,
+      "learning_rate": 9.989611685206881e-06,
+      "loss": 2.1478,
+      "step": 40
+    },
+    {
+      "epoch": 0.28975265017667845,
+      "grad_norm": 0.65234375,
+      "learning_rate": 9.987637860920053e-06,
+      "loss": 2.0103,
+      "step": 41
+    },
+    {
+      "epoch": 0.2968197879858657,
+      "grad_norm": 0.63671875,
+      "learning_rate": 9.985492694631733e-06,
+      "loss": 1.4481,
+      "step": 42
+    },
+    {
+      "epoch": 0.303886925795053,
+      "grad_norm": 0.80078125,
+      "learning_rate": 9.983176260035546e-06,
+      "loss": 1.7512,
+      "step": 43
+    },
+    {
+      "epoch": 0.31095406360424027,
+      "grad_norm": 0.71875,
+      "learning_rate": 9.980688636708744e-06,
+      "loss": 2.061,
+      "step": 44
+    },
+    {
+      "epoch": 0.31802120141342755,
+      "grad_norm": 0.59765625,
+      "learning_rate": 9.978029910109491e-06,
+      "loss": 1.572,
+      "step": 45
+    },
+    {
+      "epoch": 0.3250883392226148,
+      "grad_norm": 0.66015625,
+      "learning_rate": 9.975200171573917e-06,
+      "loss": 2.0828,
+      "step": 46
+    },
+    {
+      "epoch": 0.3321554770318021,
+      "grad_norm": 0.63671875,
+      "learning_rate": 9.972199518312979e-06,
+      "loss": 1.8459,
+      "step": 47
+    },
+    {
+      "epoch": 0.3392226148409894,
+      "grad_norm": 0.6796875,
+      "learning_rate": 9.969028053409131e-06,
+      "loss": 2.2307,
+      "step": 48
+    },
+    {
+      "epoch": 0.3462897526501767,
+      "grad_norm": 0.65625,
+      "learning_rate": 9.965685885812773e-06,
+      "loss": 1.5793,
+      "step": 49
+    },
+    {
+      "epoch": 0.35335689045936397,
+      "grad_norm": 0.75,
+      "learning_rate": 9.96217313033851e-06,
+      "loss": 2.1297,
+      "step": 50
+    },
+    {
+      "epoch": 0.36042402826855124,
+      "grad_norm": 1.0625,
+      "learning_rate": 9.958489907661217e-06,
+      "loss": 1.912,
+      "step": 51
+    },
+    {
+      "epoch": 0.3674911660777385,
+      "grad_norm": 0.72265625,
+      "learning_rate": 9.95463634431188e-06,
+      "loss": 2.122,
+      "step": 52
+    },
+    {
+      "epoch": 0.3745583038869258,
+      "grad_norm": 0.59375,
+      "learning_rate": 9.950612572673255e-06,
+      "loss": 1.9928,
+      "step": 53
+    },
+    {
+      "epoch": 0.38162544169611307,
+      "grad_norm": 0.8984375,
+      "learning_rate": 9.946418730975326e-06,
+      "loss": 1.9401,
+      "step": 54
+    },
+    {
+      "epoch": 0.38869257950530034,
+      "grad_norm": 0.578125,
+      "learning_rate": 9.942054963290549e-06,
+      "loss": 1.742,
+      "step": 55
+    },
+    {
+      "epoch": 0.3957597173144876,
+      "grad_norm": 0.70703125,
+      "learning_rate": 9.9375214195289e-06,
+      "loss": 1.775,
+      "step": 56
+    },
+    {
+      "epoch": 0.4028268551236749,
+      "grad_norm": 0.64453125,
+      "learning_rate": 9.932818255432733e-06,
+      "loss": 1.9486,
+      "step": 57
+    },
+    {
+      "epoch": 0.4098939929328622,
+      "grad_norm": 0.7265625,
+      "learning_rate": 9.92794563257143e-06,
+      "loss": 1.7127,
+      "step": 58
+    },
+    {
+      "epoch": 0.4169611307420495,
+      "grad_norm": 0.66796875,
+      "learning_rate": 9.92290371833584e-06,
+      "loss": 2.1642,
+      "step": 59
+    },
+    {
+      "epoch": 0.42402826855123676,
+      "grad_norm": 0.83984375,
+      "learning_rate": 9.91769268593254e-06,
+      "loss": 1.7884,
+      "step": 60
+    },
+    {
+      "epoch": 0.43109540636042404,
+      "grad_norm": 0.68359375,
+      "learning_rate": 9.91231271437788e-06,
+      "loss": 2.1369,
+      "step": 61
+    },
+    {
+      "epoch": 0.4381625441696113,
+      "grad_norm": 0.75390625,
+      "learning_rate": 9.906763988491834e-06,
+      "loss": 2.2394,
+      "step": 62
+    },
+    {
+      "epoch": 0.4452296819787986,
+      "grad_norm": 0.5859375,
+      "learning_rate": 9.901046698891648e-06,
+      "loss": 1.4972,
+      "step": 63
+    },
+    {
+      "epoch": 0.45229681978798586,
+      "grad_norm": 0.96875,
+      "learning_rate": 9.895161041985295e-06,
+      "loss": 1.8555,
+      "step": 64
+    },
+    {
+      "epoch": 0.45936395759717313,
+      "grad_norm": 0.74609375,
+      "learning_rate": 9.889107219964726e-06,
+      "loss": 2.2285,
+      "step": 65
+    },
+    {
+      "epoch": 0.4664310954063604,
+      "grad_norm": 0.6015625,
+      "learning_rate": 9.882885440798928e-06,
+      "loss": 1.6138,
+      "step": 66
+    },
+    {
+      "epoch": 0.4734982332155477,
+      "grad_norm": 0.6171875,
+      "learning_rate": 9.876495918226772e-06,
+      "loss": 1.7915,
+      "step": 67
+    },
+    {
+      "epoch": 0.48056537102473496,
+      "grad_norm": 0.56640625,
+      "learning_rate": 9.869938871749676e-06,
+      "loss": 1.6075,
+      "step": 68
+    },
+    {
+      "epoch": 0.4876325088339223,
+      "grad_norm": 0.6640625,
+      "learning_rate": 9.863214526624065e-06,
+      "loss": 1.9207,
+      "step": 69
+    },
+    {
+      "epoch": 0.49469964664310956,
+      "grad_norm": 0.703125,
+      "learning_rate": 9.856323113853632e-06,
+      "loss": 1.9285,
+      "step": 70
+    },
+    {
+      "epoch": 0.5017667844522968,
+      "grad_norm": 0.68359375,
+      "learning_rate": 9.849264870181393e-06,
+      "loss": 1.9508,
+      "step": 71
+    },
+    {
+      "epoch": 0.508833922261484,
+      "grad_norm": 0.6640625,
+      "learning_rate": 9.842040038081572e-06,
+      "loss": 1.9983,
+      "step": 72
+    },
+    {
+      "epoch": 0.5159010600706714,
+      "grad_norm": 0.5703125,
+      "learning_rate": 9.834648865751254e-06,
+      "loss": 1.6503,
+      "step": 73
+    },
+    {
+      "epoch": 0.5229681978798587,
+      "grad_norm": 1.046875,
+      "learning_rate": 9.827091607101871e-06,
+      "loss": 1.8562,
+      "step": 74
+    },
+    {
+      "epoch": 0.5300353356890459,
+      "grad_norm": 0.78515625,
+      "learning_rate": 9.81936852175047e-06,
+      "loss": 1.5611,
+      "step": 75
+    },
+    {
+      "epoch": 0.5371024734982333,
+      "grad_norm": 0.61328125,
+      "learning_rate": 9.811479875010801e-06,
+      "loss": 1.7828,
+      "step": 76
+    },
+    {
+      "epoch": 0.5441696113074205,
+      "grad_norm": 0.6015625,
+      "learning_rate": 9.803425937884202e-06,
+      "loss": 1.5878,
+      "step": 77
+    },
+    {
+      "epoch": 0.5512367491166078,
+      "grad_norm": 0.76171875,
+      "learning_rate": 9.79520698705028e-06,
+      "loss": 2.0218,
+      "step": 78
+    },
+    {
+      "epoch": 0.558303886925795,
+      "grad_norm": 0.640625,
+      "learning_rate": 9.786823304857424e-06,
+      "loss": 1.4419,
+      "step": 79
+    },
+    {
+      "epoch": 0.5653710247349824,
+      "grad_norm": 1.2265625,
+      "learning_rate": 9.778275179313084e-06,
+      "loss": 1.9871,
+      "step": 80
+    },
+    {
+      "epoch": 0.5724381625441696,
+      "grad_norm": 0.71875,
+      "learning_rate": 9.769562904073896e-06,
+      "loss": 1.9268,
+      "step": 81
+    },
+    {
+      "epoch": 0.5795053003533569,
+      "grad_norm": 0.57421875,
+      "learning_rate": 9.76068677843558e-06,
+      "loss": 1.4848,
+      "step": 82
+    },
+    {
+      "epoch": 0.5865724381625441,
+      "grad_norm": 0.65234375,
+      "learning_rate": 9.751647107322668e-06,
+      "loss": 2.0959,
+      "step": 83
+    },
+    {
+      "epoch": 0.5936395759717314,
+      "grad_norm": 0.74609375,
+      "learning_rate": 9.742444201278022e-06,
+      "loss": 2.3728,
+      "step": 84
+    },
+    {
+      "epoch": 0.6007067137809188,
+      "grad_norm": 0.671875,
+      "learning_rate": 9.733078376452172e-06,
+      "loss": 1.3637,
+      "step": 85
+    },
+    {
+      "epoch": 0.607773851590106,
+      "grad_norm": 0.671875,
+      "learning_rate": 9.723549954592447e-06,
+      "loss": 1.8211,
+      "step": 86
+    },
+    {
+      "epoch": 0.6148409893992933,
+      "grad_norm": 0.64453125,
+      "learning_rate": 9.713859263031928e-06,
+      "loss": 1.9919,
+      "step": 87
+    },
+    {
+      "epoch": 0.6219081272084805,
+      "grad_norm": 0.6015625,
+      "learning_rate": 9.704006634678205e-06,
+      "loss": 1.5786,
+      "step": 88
+    },
+    {
+      "epoch": 0.6289752650176679,
+      "grad_norm": 0.6640625,
+      "learning_rate": 9.693992408001934e-06,
+      "loss": 1.9939,
+      "step": 89
+    },
+    {
+      "epoch": 0.6360424028268551,
+      "grad_norm": 0.60546875,
+      "learning_rate": 9.683816927025212e-06,
+      "loss": 2.1518,
+      "step": 90
+    },
+    {
+      "epoch": 0.6431095406360424,
+      "grad_norm": 0.7578125,
+      "learning_rate": 9.673480541309761e-06,
+      "loss": 2.2444,
+      "step": 91
+    },
+    {
+      "epoch": 0.6501766784452296,
+      "grad_norm": 0.609375,
+      "learning_rate": 9.662983605944918e-06,
+      "loss": 1.9037,
+      "step": 92
+    },
+    {
+      "epoch": 0.657243816254417,
+      "grad_norm": 1.5625,
+      "learning_rate": 9.652326481535434e-06,
+      "loss": 2.3459,
+      "step": 93
+    },
+    {
+      "epoch": 0.6643109540636042,
+      "grad_norm": 0.609375,
+      "learning_rate": 9.64150953418909e-06,
+      "loss": 1.7578,
+      "step": 94
+    },
+    {
+      "epoch": 0.6713780918727915,
+      "grad_norm": 0.6328125,
+      "learning_rate": 9.630533135504118e-06,
+      "loss": 1.5717,
+      "step": 95
+    },
+    {
+      "epoch": 0.6784452296819788,
+      "grad_norm": 0.7109375,
+      "learning_rate": 9.619397662556434e-06,
+      "loss": 1.7722,
+      "step": 96
+    },
+    {
+      "epoch": 0.6855123674911661,
+      "grad_norm": 0.6796875,
+      "learning_rate": 9.608103497886687e-06,
+      "loss": 1.8594,
+      "step": 97
+    },
+    {
+      "epoch": 0.6925795053003534,
+      "grad_norm": 0.66796875,
+      "learning_rate": 9.596651029487116e-06,
+      "loss": 2.1158,
+      "step": 98
+    },
+    {
+      "epoch": 0.6996466431095406,
+      "grad_norm": 0.75,
+      "learning_rate": 9.585040650788222e-06,
+      "loss": 1.2966,
+      "step": 99
+    },
+    {
+      "epoch": 0.7067137809187279,
+      "grad_norm": 1.2890625,
+      "learning_rate": 9.573272760645248e-06,
+      "loss": 2.3551,
+      "step": 100
+    },
+    {
+      "epoch": 0.7137809187279152,
+      "grad_norm": 0.61328125,
+      "learning_rate": 9.561347763324484e-06,
+      "loss": 2.0097,
+      "step": 101
+    },
+    {
+      "epoch": 0.7208480565371025,
+      "grad_norm": 0.6640625,
+      "learning_rate": 9.549266068489377e-06,
+      "loss": 1.7785,
+      "step": 102
+    },
+    {
+      "epoch": 0.7279151943462897,
+      "grad_norm": 0.59765625,
+      "learning_rate": 9.537028091186453e-06,
+      "loss": 1.7432,
+      "step": 103
+    },
+    {
+      "epoch": 0.734982332155477,
+      "grad_norm": 0.57421875,
+      "learning_rate": 9.524634251831064e-06,
+      "loss": 1.9585,
+      "step": 104
+    },
+    {
+      "epoch": 0.7420494699646644,
+      "grad_norm": 0.7890625,
+      "learning_rate": 9.512084976192944e-06,
+      "loss": 1.8571,
+      "step": 105
+    },
+    {
+      "epoch": 0.7491166077738516,
+      "grad_norm": 0.62890625,
+      "learning_rate": 9.499380695381577e-06,
+      "loss": 1.8799,
+      "step": 106
+    },
+    {
+      "epoch": 0.7561837455830389,
+      "grad_norm": 0.7421875,
+      "learning_rate": 9.486521845831403e-06,
+      "loss": 1.8795,
+      "step": 107
+    },
+    {
+      "epoch": 0.7632508833922261,
+      "grad_norm": 0.69140625,
+      "learning_rate": 9.4735088692868e-06,
+      "loss": 1.8043,
+      "step": 108
+    },
+    {
+      "epoch": 0.7703180212014135,
+      "grad_norm": 0.8125,
+      "learning_rate": 9.460342212786933e-06,
+      "loss": 2.1479,
+      "step": 109
+    },
+    {
+      "epoch": 0.7773851590106007,
+      "grad_norm": 0.7890625,
+      "learning_rate": 9.447022328650382e-06,
+      "loss": 1.7199,
+      "step": 110
+    },
+    {
+      "epoch": 0.784452296819788,
+      "grad_norm": 0.80078125,
+      "learning_rate": 9.43354967445961e-06,
+      "loss": 2.1058,
+      "step": 111
+    },
+    {
+      "epoch": 0.7915194346289752,
+      "grad_norm": 0.828125,
+      "learning_rate": 9.419924713045234e-06,
+      "loss": 1.9554,
+      "step": 112
+    },
+    {
+      "epoch": 0.7985865724381626,
+      "grad_norm": 0.64453125,
+      "learning_rate": 9.406147912470142e-06,
+      "loss": 1.3284,
+      "step": 113
+    },
+    {
+      "epoch": 0.8056537102473498,
+      "grad_norm": 0.66015625,
+      "learning_rate": 9.392219746013399e-06,
+      "loss": 1.8721,
+      "step": 114
+    },
+    {
+      "epoch": 0.8127208480565371,
+      "grad_norm": 0.5703125,
+      "learning_rate": 9.378140692153991e-06,
+      "loss": 1.8162,
+      "step": 115
+    },
+    {
+      "epoch": 0.8197879858657244,
+      "grad_norm": 0.6171875,
+      "learning_rate": 9.363911234554394e-06,
+      "loss": 2.1554,
+      "step": 116
+    },
+    {
+      "epoch": 0.8268551236749117,
+      "grad_norm": 0.703125,
+      "learning_rate": 9.349531862043952e-06,
+      "loss": 2.1176,
+      "step": 117
+    },
+    {
+      "epoch": 0.833922261484099,
+      "grad_norm": 0.609375,
+      "learning_rate": 9.335003068602087e-06,
+      "loss": 1.7494,
+      "step": 118
+    },
+    {
+      "epoch": 0.8409893992932862,
+      "grad_norm": 0.62109375,
+      "learning_rate": 9.320325353341325e-06,
+      "loss": 1.8479,
+      "step": 119
+    },
+    {
+      "epoch": 0.8480565371024735,
+      "grad_norm": 0.6328125,
+      "learning_rate": 9.305499220490162e-06,
+      "loss": 1.5816,
+      "step": 120
+    },
+    {
+      "epoch": 0.8551236749116607,
+      "grad_norm": 0.70703125,
+      "learning_rate": 9.290525179375722e-06,
+      "loss": 2.1121,
+      "step": 121
+    },
+    {
+      "epoch": 0.8621908127208481,
+      "grad_norm": 0.67578125,
+      "learning_rate": 9.275403744406282e-06,
+      "loss": 1.9516,
+      "step": 122
+    },
+    {
+      "epoch": 0.8692579505300353,
+      "grad_norm": 0.609375,
+      "learning_rate": 9.260135435053583e-06,
+      "loss": 1.8081,
+      "step": 123
+    },
+    {
+      "epoch": 0.8763250883392226,
+      "grad_norm": 0.62890625,
+      "learning_rate": 9.244720775834993e-06,
+      "loss": 2.2066,
+      "step": 124
+    },
+    {
+      "epoch": 0.8833922261484098,
+      "grad_norm": 0.80859375,
+      "learning_rate": 9.229160296295488e-06,
+      "loss": 2.0408,
+      "step": 125
+    },
+    {
+      "epoch": 0.8904593639575972,
+      "grad_norm": 0.6875,
+      "learning_rate": 9.213454530989454e-06,
+      "loss": 1.7211,
+      "step": 126
+    },
+    {
+      "epoch": 0.8975265017667845,
+      "grad_norm": 0.58984375,
+      "learning_rate": 9.197604019462334e-06,
+      "loss": 0.905,
+      "step": 127
+    },
+    {
+      "epoch": 0.9045936395759717,
+      "grad_norm": 0.65234375,
+      "learning_rate": 9.18160930623208e-06,
+      "loss": 1.4443,
+      "step": 128
+    },
+    {
+      "epoch": 0.911660777385159,
+      "grad_norm": 0.72265625,
+      "learning_rate": 9.165470940770458e-06,
+      "loss": 1.8498,
+      "step": 129
+    },
+    {
+      "epoch": 0.9187279151943463,
+      "grad_norm": 0.6875,
+      "learning_rate": 9.149189477484169e-06,
+      "loss": 1.7645,
+      "step": 130
+    },
+    {
+      "epoch": 0.9257950530035336,
+      "grad_norm": 0.59375,
+      "learning_rate": 9.132765475695795e-06,
+      "loss": 1.7509,
+      "step": 131
+    },
+    {
+      "epoch": 0.9328621908127208,
+      "grad_norm": 0.61328125,
+      "learning_rate": 9.116199499624596e-06,
+      "loss": 2.2509,
+      "step": 132
+    },
+    {
+      "epoch": 0.9399293286219081,
+      "grad_norm": 0.765625,
+      "learning_rate": 9.099492118367123e-06,
+      "loss": 1.5412,
+      "step": 133
+    },
+    {
+      "epoch": 0.9469964664310954,
+      "grad_norm": 0.609375,
+      "learning_rate": 9.082643905877665e-06,
+      "loss": 1.5065,
+      "step": 134
+    },
+    {
+      "epoch": 0.9540636042402827,
+      "grad_norm": 0.625,
+      "learning_rate": 9.065655440948536e-06,
+      "loss": 1.8347,
+      "step": 135
+    },
+    {
+      "epoch": 0.9611307420494699,
+      "grad_norm": 0.765625,
+      "learning_rate": 9.048527307190182e-06,
+      "loss": 2.1884,
+      "step": 136
+    },
+    {
+      "epoch": 0.9681978798586572,
+      "grad_norm": 0.62109375,
+      "learning_rate": 9.03126009301115e-06,
+      "loss": 1.7727,
+      "step": 137
+    },
+    {
+      "epoch": 0.9752650176678446,
+      "grad_norm": 0.61328125,
+      "learning_rate": 9.013854391597856e-06,
+      "loss": 1.7151,
+      "step": 138
+    },
+    {
+      "epoch": 0.9823321554770318,
+      "grad_norm": 0.66796875,
+      "learning_rate": 8.996310800894215e-06,
+      "loss": 1.992,
+      "step": 139
+    },
+    {
+      "epoch": 0.9893992932862191,
+      "grad_norm": 0.69140625,
+      "learning_rate": 8.978629923581104e-06,
+      "loss": 2.2281,
+      "step": 140
+    },
+    {
+      "epoch": 0.9964664310954063,
+      "grad_norm": 1.15625,
+      "learning_rate": 8.960812367055646e-06,
+      "loss": 2.874,
+      "step": 141
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 1.0625,
+      "learning_rate": 8.94285874341036e-06,
+      "loss": 2.0683,
+      "step": 142
+    },
+    {
+      "epoch": 1.0070671378091873,
+      "grad_norm": 0.59375,
+      "learning_rate": 8.924769669412117e-06,
+      "loss": 1.8785,
+      "step": 143
+    },
+    {
+      "epoch": 1.0141342756183747,
+      "grad_norm": 0.46484375,
+      "learning_rate": 8.906545766480961e-06,
+      "loss": 1.1151,
+      "step": 144
+    },
+    {
+      "epoch": 1.0212014134275618,
+      "grad_norm": 0.55859375,
+      "learning_rate": 8.888187660668762e-06,
+      "loss": 1.6426,
+      "step": 145
+    },
+    {
+      "epoch": 1.028268551236749,
+      "grad_norm": 0.57421875,
+      "learning_rate": 8.869695982637703e-06,
+      "loss": 1.5826,
+      "step": 146
+    },
+    {
+      "epoch": 1.0353356890459364,
+      "grad_norm": 0.6171875,
+      "learning_rate": 8.851071367638625e-06,
+      "loss": 1.844,
+      "step": 147
+    },
+    {
+      "epoch": 1.0424028268551238,
+      "grad_norm": 1.03125,
+      "learning_rate": 8.832314455489188e-06,
+      "loss": 1.8547,
+      "step": 148
+    },
+    {
+      "epoch": 1.0494699646643109,
+      "grad_norm": 0.6796875,
+      "learning_rate": 8.81342589055191e-06,
+      "loss": 1.834,
+      "step": 149
+    },
+    {
+      "epoch": 1.0565371024734982,
+      "grad_norm": 0.64453125,
+      "learning_rate": 8.794406321712017e-06,
+      "loss": 1.7167,
+      "step": 150
+    },
+    {
+      "epoch": 1.0636042402826855,
+      "grad_norm": 0.55859375,
+      "learning_rate": 8.775256402355155e-06,
+      "loss": 1.6672,
+      "step": 151
+    },
+    {
+      "epoch": 1.0706713780918728,
+      "grad_norm": 0.68359375,
+      "learning_rate": 8.755976790344945e-06,
+      "loss": 1.7684,
+      "step": 152
+    },
+    {
+      "epoch": 1.0777385159010602,
+      "grad_norm": 0.609375,
+      "learning_rate": 8.736568148000386e-06,
+      "loss": 2.1258,
+      "step": 153
+    },
+    {
+      "epoch": 1.0848056537102473,
+      "grad_norm": 0.65234375,
+      "learning_rate": 8.7170311420731e-06,
+      "loss": 1.8917,
+      "step": 154
+    },
+    {
+      "epoch": 1.0918727915194346,
+      "grad_norm": 0.93359375,
+      "learning_rate": 8.697366443724424e-06,
+      "loss": 2.4136,
+      "step": 155
+    },
+    {
+      "epoch": 1.098939929328622,
+      "grad_norm": 0.61328125,
+      "learning_rate": 8.677574728502355e-06,
+      "loss": 1.9583,
+      "step": 156
+    },
+    {
+      "epoch": 1.1060070671378093,
+      "grad_norm": 0.62109375,
+      "learning_rate": 8.657656676318346e-06,
+      "loss": 1.974,
+      "step": 157
+    },
+    {
+      "epoch": 1.1130742049469964,
+      "grad_norm": 0.58203125,
+      "learning_rate": 8.637612971423945e-06,
+      "loss": 1.4882,
+      "step": 158
+    },
+    {
+      "epoch": 1.1201413427561837,
+      "grad_norm": 0.61328125,
+      "learning_rate": 8.617444302387288e-06,
+      "loss": 2.2903,
+      "step": 159
+    },
+    {
+      "epoch": 1.127208480565371,
+      "grad_norm": 0.55078125,
+      "learning_rate": 8.597151362069452e-06,
+      "loss": 1.8166,
+      "step": 160
+    },
+    {
+      "epoch": 1.1342756183745584,
+      "grad_norm": 0.578125,
+      "learning_rate": 8.576734847600639e-06,
+      "loss": 1.9671,
+      "step": 161
+    },
+    {
+      "epoch": 1.1413427561837457,
+      "grad_norm": 0.53125,
+      "learning_rate": 8.55619546035624e-06,
+      "loss": 1.6708,
+      "step": 162
+    },
+    {
+      "epoch": 1.1484098939929328,
+      "grad_norm": 1.2109375,
+      "learning_rate": 8.535533905932739e-06,
+      "loss": 2.8918,
+      "step": 163
+    },
+    {
+      "epoch": 1.1554770318021201,
+      "grad_norm": 0.57421875,
+      "learning_rate": 8.514750894123463e-06,
+      "loss": 1.6384,
+      "step": 164
+    },
+    {
+      "epoch": 1.1625441696113075,
+      "grad_norm": 0.62109375,
+      "learning_rate": 8.49384713889421e-06,
+      "loss": 1.8558,
+      "step": 165
+    },
+    {
+      "epoch": 1.1696113074204948,
+      "grad_norm": 0.5,
+      "learning_rate": 8.472823358358716e-06,
+      "loss": 1.5123,
+      "step": 166
+    },
+    {
+      "epoch": 1.176678445229682,
+      "grad_norm": 0.6171875,
+      "learning_rate": 8.451680274753986e-06,
+      "loss": 2.0303,
+      "step": 167
+    },
+    {
+      "epoch": 1.1837455830388692,
+      "grad_norm": 0.640625,
+      "learning_rate": 8.430418614415488e-06,
+      "loss": 1.8954,
+      "step": 168
+    },
+    {
+      "epoch": 1.1908127208480566,
+      "grad_norm": 0.609375,
+      "learning_rate": 8.40903910775219e-06,
+      "loss": 2.1229,
+      "step": 169
+    },
+    {
+      "epoch": 1.197879858657244,
+      "grad_norm": 0.6796875,
+      "learning_rate": 8.387542489221477e-06,
+      "loss": 1.6479,
+      "step": 170
+    },
+    {
+      "epoch": 1.2049469964664312,
+      "grad_norm": 0.68359375,
+      "learning_rate": 8.36592949730392e-06,
+      "loss": 1.9559,
+      "step": 171
+    },
+    {
+      "epoch": 1.2120141342756183,
+      "grad_norm": 0.921875,
+      "learning_rate": 8.344200874477901e-06,
+      "loss": 2.436,
+      "step": 172
+    },
+    {
+      "epoch": 1.2190812720848057,
+      "grad_norm": 0.64453125,
+      "learning_rate": 8.32235736719411e-06,
+      "loss": 1.8343,
+      "step": 173
+    },
+    {
+      "epoch": 1.226148409893993,
+      "grad_norm": 0.61328125,
+      "learning_rate": 8.300399725849902e-06,
+      "loss": 2.1196,
+      "step": 174
+    },
+    {
+      "epoch": 1.23321554770318,
+      "grad_norm": 0.6953125,
+      "learning_rate": 8.278328704763516e-06,
+      "loss": 2.2121,
+      "step": 175
+    },
+    {
+      "epoch": 1.2402826855123674,
+      "grad_norm": 0.703125,
+      "learning_rate": 8.256145062148168e-06,
+      "loss": 1.8644,
+      "step": 176
+    },
+    {
+      "epoch": 1.2473498233215548,
+      "grad_norm": 0.67578125,
+      "learning_rate": 8.233849560085994e-06,
+      "loss": 2.0178,
+      "step": 177
+    },
+    {
+      "epoch": 1.254416961130742,
+      "grad_norm": 0.6640625,
+      "learning_rate": 8.211442964501879e-06,
+      "loss": 2.1933,
+      "step": 178
+    },
+    {
+      "epoch": 1.2614840989399294,
+      "grad_norm": 0.63671875,
+      "learning_rate": 8.188926045137139e-06,
+      "loss": 1.9427,
+      "step": 179
+    },
+    {
+      "epoch": 1.2685512367491167,
+      "grad_norm": 0.6640625,
+      "learning_rate": 8.166299575523081e-06,
+      "loss": 1.9772,
+      "step": 180
+    },
+    {
+      "epoch": 1.2756183745583038,
+      "grad_norm": 0.6484375,
+      "learning_rate": 8.143564332954426e-06,
+      "loss": 1.6406,
+      "step": 181
+    },
+    {
+      "epoch": 1.2826855123674912,
+      "grad_norm": 0.60546875,
+      "learning_rate": 8.120721098462612e-06,
+      "loss": 1.9615,
+      "step": 182
+    },
+    {
+      "epoch": 1.2897526501766785,
+      "grad_norm": 0.68359375,
+      "learning_rate": 8.097770656788961e-06,
+      "loss": 1.9663,
+      "step": 183
+    },
+    {
+      "epoch": 1.2968197879858656,
+      "grad_norm": 0.62109375,
+      "learning_rate": 8.074713796357717e-06,
+      "loss": 1.756,
+      "step": 184
+    },
+    {
+      "epoch": 1.303886925795053,
+      "grad_norm": 0.69140625,
+      "learning_rate": 8.051551309248961e-06,
+      "loss": 1.8254,
+      "step": 185
+    },
+    {
+      "epoch": 1.3109540636042403,
+      "grad_norm": 0.66015625,
+      "learning_rate": 8.028283991171408e-06,
+      "loss": 1.9618,
+      "step": 186
+    },
+    {
+      "epoch": 1.3180212014134276,
+      "grad_norm": 0.6484375,
+      "learning_rate": 8.004912641435064e-06,
+      "loss": 1.9003,
+      "step": 187
+    },
+    {
+      "epoch": 1.325088339222615,
+      "grad_norm": 0.6640625,
+      "learning_rate": 7.981438062923767e-06,
+      "loss": 1.9284,
+      "step": 188
+    },
+    {
+      "epoch": 1.332155477031802,
+      "grad_norm": 0.71484375,
+      "learning_rate": 7.957861062067614e-06,
+      "loss": 1.9229,
+      "step": 189
+    },
+    {
+      "epoch": 1.3392226148409894,
+      "grad_norm": 0.703125,
+      "learning_rate": 7.934182448815244e-06,
+      "loss": 1.5828,
+      "step": 190
+    },
+    {
+      "epoch": 1.3462897526501767,
+      "grad_norm": 0.58203125,
+      "learning_rate": 7.910403036606028e-06,
+      "loss": 2.1854,
+      "step": 191
+    },
+    {
+      "epoch": 1.353356890459364,
+      "grad_norm": 0.55859375,
+      "learning_rate": 7.88652364234211e-06,
+      "loss": 1.492,
+      "step": 192
+    },
+    {
+      "epoch": 1.3604240282685511,
+      "grad_norm": 0.625,
+      "learning_rate": 7.86254508636036e-06,
+      "loss": 1.6163,
+      "step": 193
+    },
+    {
+      "epoch": 1.3674911660777385,
+      "grad_norm": 0.73046875,
+      "learning_rate": 7.838468192404176e-06,
+      "loss": 1.8491,
+      "step": 194
+    },
+    {
+      "epoch": 1.3745583038869258,
+      "grad_norm": 0.59765625,
+      "learning_rate": 7.814293787595197e-06,
+      "loss": 1.472,
+      "step": 195
+    },
+    {
+      "epoch": 1.3816254416961131,
+      "grad_norm": 0.62890625,
+      "learning_rate": 7.790022702404887e-06,
+      "loss": 2.0314,
+      "step": 196
+    },
+    {
+      "epoch": 1.3886925795053005,
+      "grad_norm": 0.64453125,
+      "learning_rate": 7.765655770625997e-06,
+      "loss": 1.7422,
+      "step": 197
+    },
+    {
+      "epoch": 1.3957597173144876,
+      "grad_norm": 0.625,
+      "learning_rate": 7.741193829343937e-06,
+      "loss": 1.1329,
+      "step": 198
+    },
+    {
+      "epoch": 1.4028268551236749,
+      "grad_norm": 0.9140625,
+      "learning_rate": 7.716637718908002e-06,
+      "loss": 1.4366,
+      "step": 199
+    },
+    {
+      "epoch": 1.4098939929328622,
+      "grad_norm": 0.56640625,
+      "learning_rate": 7.691988282902519e-06,
+      "loss": 1.4888,
+      "step": 200
+    },
+    {
+      "epoch": 1.4169611307420495,
+      "grad_norm": 0.69140625,
+      "learning_rate": 7.667246368117852e-06,
+      "loss": 1.8835,
+      "step": 201
+    },
+    {
+      "epoch": 1.4240282685512367,
+      "grad_norm": 0.60546875,
+      "learning_rate": 7.642412824521328e-06,
+      "loss": 1.6722,
+      "step": 202
+    },
+    {
+      "epoch": 1.431095406360424,
+      "grad_norm": 0.6328125,
+      "learning_rate": 7.617488505228023e-06,
+      "loss": 2.0471,
+      "step": 203
+    },
+    {
+      "epoch": 1.4381625441696113,
+      "grad_norm": 0.58203125,
+      "learning_rate": 7.592474266471464e-06,
+      "loss": 1.944,
+      "step": 204
+    },
+    {
+      "epoch": 1.4452296819787986,
+      "grad_norm": 0.6171875,
+      "learning_rate": 7.56737096757421e-06,
+      "loss": 1.6996,
+      "step": 205
+    },
+    {
+      "epoch": 1.452296819787986,
+      "grad_norm": 0.7109375,
+      "learning_rate": 7.542179470918336e-06,
+      "loss": 1.898,
+      "step": 206
+    },
+    {
+      "epoch": 1.459363957597173,
+      "grad_norm": 0.62890625,
+      "learning_rate": 7.5169006419157985e-06,
+      "loss": 1.6575,
+      "step": 207
+    },
+    {
+      "epoch": 1.4664310954063604,
+      "grad_norm": 0.6328125,
+      "learning_rate": 7.491535348978719e-06,
+      "loss": 1.6194,
+      "step": 208
+    },
+    {
+      "epoch": 1.4734982332155477,
+      "grad_norm": 0.79296875,
+      "learning_rate": 7.466084463489537e-06,
+      "loss": 2.1004,
+      "step": 209
+    },
+    {
+      "epoch": 1.4805653710247348,
+      "grad_norm": 0.8828125,
+      "learning_rate": 7.440548859771086e-06,
+      "loss": 2.1903,
+      "step": 210
+    },
+    {
+      "epoch": 1.4876325088339222,
+      "grad_norm": 0.73046875,
+      "learning_rate": 7.414929415056551e-06,
+      "loss": 2.0521,
+      "step": 211
+    },
+    {
+      "epoch": 1.4946996466431095,
+      "grad_norm": 0.6484375,
+      "learning_rate": 7.389227009459335e-06,
+      "loss": 1.8102,
+      "step": 212
+    },
+    {
+      "epoch": 1.5017667844522968,
+      "grad_norm": 0.56640625,
+      "learning_rate": 7.363442525942827e-06,
+      "loss": 1.8402,
+      "step": 213
+    },
+    {
+      "epoch": 1.5088339222614842,
+      "grad_norm": 0.65234375,
+      "learning_rate": 7.3375768502900626e-06,
+      "loss": 1.949,
+      "step": 214
+    },
+    {
+      "epoch": 1.5159010600706715,
+      "grad_norm": 0.65625,
+      "learning_rate": 7.311630871073301e-06,
+      "loss": 1.6098,
+      "step": 215
+    },
+    {
+      "epoch": 1.5229681978798588,
+      "grad_norm": 0.62890625,
+      "learning_rate": 7.2856054796234944e-06,
+      "loss": 1.4942,
+      "step": 216
+    },
+    {
+      "epoch": 1.530035335689046,
+      "grad_norm": 0.6328125,
+      "learning_rate": 7.25950156999967e-06,
+      "loss": 1.9404,
+      "step": 217
+    },
+    {
+      "epoch": 1.5371024734982333,
+      "grad_norm": 0.64453125,
+      "learning_rate": 7.23332003895822e-06,
+      "loss": 1.8286,
+      "step": 218
+    },
+    {
+      "epoch": 1.5441696113074204,
+      "grad_norm": 0.75390625,
+      "learning_rate": 7.207061785922089e-06,
+      "loss": 2.0283,
+      "step": 219
+    },
+    {
+      "epoch": 1.5512367491166077,
+      "grad_norm": 0.6953125,
+      "learning_rate": 7.1807277129498774e-06,
+      "loss": 1.8301,
+      "step": 220
+    },
+    {
+      "epoch": 1.558303886925795,
+      "grad_norm": 0.62109375,
+      "learning_rate": 7.1543187247048525e-06,
+      "loss": 2.0651,
+      "step": 221
+    },
+    {
+      "epoch": 1.5653710247349824,
+      "grad_norm": 0.6015625,
+      "learning_rate": 7.1278357284238745e-06,
+      "loss": 1.6387,
+      "step": 222
+    },
+    {
+      "epoch": 1.5724381625441697,
+      "grad_norm": 0.64453125,
+      "learning_rate": 7.101279633886222e-06,
+      "loss": 1.5276,
+      "step": 223
+    },
+    {
+      "epoch": 1.579505300353357,
+      "grad_norm": 0.609375,
+      "learning_rate": 7.074651353382349e-06,
+      "loss": 1.8464,
+      "step": 224
+    },
+    {
+      "epoch": 1.5865724381625441,
+      "grad_norm": 0.65625,
+      "learning_rate": 7.047951801682533e-06,
+      "loss": 2.1482,
+      "step": 225
+    },
+    {
+      "epoch": 1.5936395759717314,
+      "grad_norm": 0.63671875,
+      "learning_rate": 7.021181896005456e-06,
+      "loss": 2.2091,
+      "step": 226
+    },
+    {
+      "epoch": 1.6007067137809188,
+      "grad_norm": 0.59765625,
+      "learning_rate": 6.994342555986692e-06,
+      "loss": 1.707,
+      "step": 227
+    },
+    {
+      "epoch": 1.6077738515901059,
+      "grad_norm": 0.6015625,
+      "learning_rate": 6.967434703647123e-06,
+      "loss": 1.7147,
+      "step": 228
+    },
+    {
+      "epoch": 1.6148409893992932,
+      "grad_norm": 0.59375,
+      "learning_rate": 6.9404592633612486e-06,
+      "loss": 1.5438,
+      "step": 229
+    },
+    {
+      "epoch": 1.6219081272084805,
+      "grad_norm": 0.640625,
+      "learning_rate": 6.913417161825449e-06,
+      "loss": 1.8477,
+      "step": 230
+    },
+    {
+      "epoch": 1.6289752650176679,
+      "grad_norm": 0.55859375,
+      "learning_rate": 6.886309328026135e-06,
+      "loss": 1.4667,
+      "step": 231
+    },
+    {
+      "epoch": 1.6360424028268552,
+      "grad_norm": 0.5859375,
+      "learning_rate": 6.85913669320784e-06,
+      "loss": 1.4187,
+      "step": 232
+    },
+    {
+      "epoch": 1.6431095406360425,
+      "grad_norm": 0.59765625,
+      "learning_rate": 6.831900190841232e-06,
+      "loss": 2.2355,
+      "step": 233
+    },
+    {
+      "epoch": 1.6501766784452296,
+      "grad_norm": 0.5703125,
+      "learning_rate": 6.804600756591037e-06,
+      "loss": 1.9369,
+      "step": 234
+    },
+    {
+      "epoch": 1.657243816254417,
+      "grad_norm": 0.671875,
+      "learning_rate": 6.777239328283909e-06,
+      "loss": 1.8834,
+      "step": 235
+    },
+    {
+      "epoch": 1.664310954063604,
+      "grad_norm": 0.65625,
+      "learning_rate": 6.749816845876196e-06,
+      "loss": 1.8316,
+      "step": 236
+    },
+    {
+      "epoch": 1.6713780918727914,
+      "grad_norm": 0.5859375,
+      "learning_rate": 6.722334251421665e-06,
+      "loss": 1.5946,
+      "step": 237
+    },
+    {
+      "epoch": 1.6784452296819787,
+      "grad_norm": 0.625,
+      "learning_rate": 6.6947924890391295e-06,
+      "loss": 1.6439,
+      "step": 238
+    },
+    {
+      "epoch": 1.685512367491166,
+      "grad_norm": 0.625,
+      "learning_rate": 6.667192504880016e-06,
+      "loss": 2.0241,
+      "step": 239
+    },
+    {
+      "epoch": 1.6925795053003534,
+      "grad_norm": 0.59375,
+      "learning_rate": 6.639535247095868e-06,
+      "loss": 1.5223,
+      "step": 240
+    },
+    {
+      "epoch": 1.6996466431095407,
+      "grad_norm": 0.6796875,
+      "learning_rate": 6.611821665805769e-06,
+      "loss": 2.1195,
+      "step": 241
+    },
+    {
+      "epoch": 1.706713780918728,
+      "grad_norm": 0.51953125,
+      "learning_rate": 6.5840527130637e-06,
+      "loss": 1.4851,
+      "step": 242
+    },
+    {
+      "epoch": 1.7137809187279152,
+      "grad_norm": 0.63671875,
+      "learning_rate": 6.556229342825835e-06,
+      "loss": 1.8208,
+      "step": 243
+    },
+    {
+      "epoch": 1.7208480565371025,
+      "grad_norm": 0.59765625,
+      "learning_rate": 6.528352510917774e-06,
+      "loss": 1.5442,
+      "step": 244
+    },
+    {
+      "epoch": 1.7279151943462896,
+      "grad_norm": 0.66015625,
+      "learning_rate": 6.500423175001705e-06,
+      "loss": 1.8611,
+      "step": 245
+    },
+    {
+      "epoch": 1.734982332155477,
+      "grad_norm": 0.953125,
+      "learning_rate": 6.472442294543497e-06,
+      "loss": 2.7323,
+      "step": 246
+    },
+    {
+      "epoch": 1.7420494699646643,
+      "grad_norm": 0.83203125,
+      "learning_rate": 6.444410830779753e-06,
+      "loss": 1.4034,
+      "step": 247
+    },
+    {
+      "epoch": 1.7491166077738516,
+      "grad_norm": 0.609375,
+      "learning_rate": 6.4163297466847795e-06,
+      "loss": 2.1511,
+      "step": 248
+    },
+    {
+      "epoch": 1.756183745583039,
+      "grad_norm": 0.60546875,
+      "learning_rate": 6.388200006937503e-06,
+      "loss": 1.6779,
+      "step": 249
+    },
+    {
+      "epoch": 1.7632508833922262,
+      "grad_norm": 0.5703125,
+      "learning_rate": 6.3600225778883395e-06,
+      "loss": 1.3638,
+      "step": 250
+    },
+    {
+      "epoch": 1.7703180212014136,
+      "grad_norm": 0.625,
+      "learning_rate": 6.33179842752599e-06,
+      "loss": 1.8632,
+      "step": 251
+    },
+    {
+      "epoch": 1.7773851590106007,
+      "grad_norm": 0.59765625,
+      "learning_rate": 6.303528525444185e-06,
+      "loss": 2.0674,
+      "step": 252
+    },
+    {
+      "epoch": 1.784452296819788,
+      "grad_norm": 0.80078125,
+      "learning_rate": 6.275213842808383e-06,
+      "loss": 2.6682,
+      "step": 253
+    },
+    {
+      "epoch": 1.7915194346289751,
+      "grad_norm": 0.90234375,
+      "learning_rate": 6.246855352322403e-06,
+      "loss": 3.1173,
+      "step": 254
+    },
+    {
+      "epoch": 1.7985865724381624,
+      "grad_norm": 0.6171875,
+      "learning_rate": 6.21845402819501e-06,
+      "loss": 2.3359,
+      "step": 255
+    },
+    {
+      "epoch": 1.8056537102473498,
+      "grad_norm": 0.59765625,
+      "learning_rate": 6.190010846106446e-06,
+      "loss": 1.6528,
+      "step": 256
+    },
+    {
+      "epoch": 1.812720848056537,
+      "grad_norm": 0.69140625,
+      "learning_rate": 6.161526783174917e-06,
+      "loss": 2.0144,
+      "step": 257
+    },
+    {
+      "epoch": 1.8197879858657244,
+      "grad_norm": 0.6796875,
+      "learning_rate": 6.1330028179230185e-06,
+      "loss": 2.078,
+      "step": 258
+    },
+    {
+      "epoch": 1.8268551236749118,
+      "grad_norm": 0.6640625,
+      "learning_rate": 6.104439930244125e-06,
+      "loss": 2.0176,
+      "step": 259
+    },
+    {
+      "epoch": 1.833922261484099,
+      "grad_norm": 0.95703125,
+      "learning_rate": 6.075839101368728e-06,
+      "loss": 1.7709,
+      "step": 260
+    },
+    {
+      "epoch": 1.8409893992932862,
+      "grad_norm": 0.6484375,
+      "learning_rate": 6.047201313830724e-06,
+      "loss": 1.7037,
+      "step": 261
+    },
+    {
+      "epoch": 1.8480565371024735,
+      "grad_norm": 0.6796875,
+      "learning_rate": 6.01852755143366e-06,
+      "loss": 1.8551,
+      "step": 262
+    },
+    {
+      "epoch": 1.8551236749116606,
+      "grad_norm": 0.640625,
+      "learning_rate": 5.98981879921695e-06,
+      "loss": 1.7625,
+      "step": 263
+    },
+    {
+      "epoch": 1.862190812720848,
+      "grad_norm": 0.59765625,
+      "learning_rate": 5.961076043422011e-06,
+      "loss": 2.0804,
+      "step": 264
+    },
+    {
+      "epoch": 1.8692579505300353,
+      "grad_norm": 0.625,
+      "learning_rate": 5.932300271458406e-06,
+      "loss": 1.7134,
+      "step": 265
+    },
+    {
+      "epoch": 1.8763250883392226,
+      "grad_norm": 0.57421875,
+      "learning_rate": 5.90349247186991e-06,
+      "loss": 1.9591,
+      "step": 266
+    },
+    {
+      "epoch": 1.88339222614841,
+      "grad_norm": 0.71875,
+      "learning_rate": 5.874653634300555e-06,
+      "loss": 1.8638,
+      "step": 267
+    },
+    {
+      "epoch": 1.8904593639575973,
+      "grad_norm": 0.65625,
+      "learning_rate": 5.845784749460632e-06,
+      "loss": 1.9674,
+      "step": 268
+    },
+    {
+      "epoch": 1.8975265017667846,
+      "grad_norm": 0.60546875,
+      "learning_rate": 5.816886809092651e-06,
+      "loss": 1.8133,
+      "step": 269
+    },
+    {
+      "epoch": 1.9045936395759717,
+      "grad_norm": 0.71484375,
+      "learning_rate": 5.787960805937283e-06,
+      "loss": 2.166,
+      "step": 270
+    },
+    {
+      "epoch": 1.911660777385159,
+      "grad_norm": 0.61328125,
+      "learning_rate": 5.759007733699245e-06,
+      "loss": 1.818,
+      "step": 271
+    },
+    {
+      "epoch": 1.9187279151943462,
+      "grad_norm": 0.58203125,
+      "learning_rate": 5.730028587013168e-06,
+      "loss": 1.4026,
+      "step": 272
+    },
+    {
+      "epoch": 1.9257950530035335,
+      "grad_norm": 0.69921875,
+      "learning_rate": 5.701024361409431e-06,
+      "loss": 2.0534,
+      "step": 273
+    },
+    {
+      "epoch": 1.9328621908127208,
+      "grad_norm": 0.734375,
+      "learning_rate": 5.671996053279949e-06,
+      "loss": 1.7194,
+      "step": 274
+    },
+    {
+      "epoch": 1.9399293286219081,
+      "grad_norm": 0.69140625,
+      "learning_rate": 5.642944659843962e-06,
+      "loss": 2.1351,
+      "step": 275
+    },
+    {
+      "epoch": 1.9469964664310955,
+      "grad_norm": 0.6796875,
+      "learning_rate": 5.613871179113761e-06,
+      "loss": 1.8866,
+      "step": 276
+    },
+    {
+      "epoch": 1.9540636042402828,
+      "grad_norm": 0.69921875,
+      "learning_rate": 5.584776609860414e-06,
+      "loss": 2.0398,
+      "step": 277
+    },
+    {
+      "epoch": 1.96113074204947,
+      "grad_norm": 0.703125,
+      "learning_rate": 5.555661951579443e-06,
+      "loss": 1.4643,
+      "step": 278
+    },
+    {
+      "epoch": 1.9681978798586572,
+      "grad_norm": 0.58984375,
+      "learning_rate": 5.5265282044565005e-06,
+      "loss": 1.4311,
+      "step": 279
+    },
+    {
+      "epoch": 1.9752650176678446,
+      "grad_norm": 0.69921875,
+      "learning_rate": 5.497376369333005e-06,
+      "loss": 1.8944,
+      "step": 280
+    },
+    {
+      "epoch": 1.9823321554770317,
+      "grad_norm": 0.62109375,
+      "learning_rate": 5.468207447671755e-06,
+      "loss": 2.3742,
+      "step": 281
+    },
+    {
+      "epoch": 1.989399293286219,
+      "grad_norm": 0.625,
+      "learning_rate": 5.43902244152253e-06,
+      "loss": 1.7038,
+      "step": 282
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 564,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 4,
+  "save_steps": 141,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 8.84019452486615e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-282/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:406982d99a22da9752c26535cc9f5894b29792d701bc3fe309a0db5d133444df
+size 6840